karashiiro/gpt-oss-20b-uwufied
karashiiro/gpt-oss-20b-uwufied is a 20 billion parameter GPT-OSS model developed by karashiiro, fine-tuned using Unsloth and Huggingface's TRL library. This model is specifically designed to generate text in an "uwu-fied" style, characterized by exaggerated cuteness and internet slang, as demonstrated in its sample outputs. It offers a 32768 token context length and is suitable for creative text generation where a distinct, playful tone is desired.
Loading preview...
Model Overview
karashiiro/gpt-oss-20b-uwufied is a 20 billion parameter language model, fine-tuned by karashiiro from the unsloth/gpt-oss-20b-unsloth-bnb-4bit base model. This fine-tuning process leveraged Unsloth for 2x faster training and Huggingface's TRL library. The model's primary distinguishing feature is its unique "uwu-fied" output style, incorporating internet slang, emojis, and a generally cutesy tone.
Key Capabilities
- "Uwu-fied" Text Generation: Produces highly stylized text with characteristic internet slang, emojis, and an exaggerated cute tone.
- Fast Fine-tuning: Benefits from Unsloth's optimization for quicker training times.
- Large Context Window: Supports a context length of 32768 tokens, allowing for processing and generating longer texts while maintaining its unique style.
Good For
- Creative Writing & Roleplay: Ideal for generating dialogue, stories, or character responses in a distinct, playful, and internet-culture-infused style.
- Novelty Applications: Suitable for projects requiring a unique and engaging textual output that stands out from standard LLM responses.
- Experimentation with Stylistic Generation: Developers interested in exploring models with highly specific and non-standard output styles will find this model useful.
Users should be aware that while the model is available in HF format, GGUF is recommended if loading issues are encountered.